inconsistent transfer prices and the location of...
TRANSCRIPT
Inconsistent Transfer Pricesand the Location of Mobile Capital
Anja De Waegenaere� Richard Sansingy
September 15, 2008
�Department of Accountancy, Tilburg University. P.O. box 90153, 5000LE Tilburg, The Nether-lands.
yTuck School of Business at Dartmouth and CentER, Tilburg University. Correspond-ing author. Tuck School of Business at Dartmouth, 100 Tuck Hall, Hanover, NH [email protected]
Inconsistent Transfer Prices and the Location of Mobile Capital
Abstract-We investigate the e¤ects of inconsistent transfer prices on the location of
capital investments made by multinational �rms in a competitive equilibrium.
Investment location depends on the tax rates that domestic and foreign investments
face. We show how transfer price inconsistency and repatriation taxes a¤ect the
combined foreign and domestic tax rate on foreign production. The tax rate on foreign
investment depends on up to �ve factors: the domestic and foreign statutory tax rates,
the domestic and foreign transfer pricing rules, and the after-tax discount rate
shareholders use to value riskless cash �ows of the �rm. We then examine the e¤ects of
inconsistent transfer prices on production decisions and productive e¢ ciency in a
model in which productive e¢ ciency and capital export neutrality are equivalent.
Eliminating the di¤erence between domestic and foreign transfer pricing rules makes
foreign investment weakly more attractive, but could either increase or decrease
productive e¢ ciency. The e¢ ciency e¤ects of changes in either the domestic or foreign
corporate tax rate, or in the shareholder after-tax discount rate, depend on the degree
of transfer price inconsistency.
1 Introduction
The allocation of income between a parent corporation and its subsidiary depends on
the price at which goods and services are transferred between the two entities. When
the entities operate in di¤erent countries, the transfer price determines how much of
the income earned by the joint e¤orts of the two entities is taxed in each country. The
determination of the transfer price used for tax purposes is di¢ cult in principle and
contentious in practice. Income earned by a multinational enterprise is subject to
double taxation to the extent that governments use inconsistent transfer prices to
allocate income between countries.
Using a model of competitive equilibrium in which all after-tax economic pro�ts
are driven to zero, we investigate how taxes a¤ect the location and e¢ ciency of capital
investments in the presence of transfer price inconsistency. First, we examine a
benchmark case that features a system of worldwide taxation, deferral of domestic
taxation of foreign source income until repatriation, and inconsistent transfer prices,
which re�ects how the U.S. currently taxes international income. Second, we
investigate the e¤ects of harmonizing transfer prices on investment location and
e¢ ciency. Third, we explore how transfer price inconsistency changes the e¤ects on
production location and productive e¢ ciency of reductions in the domestic and foreign
corporate tax rates, and of changes in the shareholder after-tax discount rate.
In our model, both countries adhere to the �arm�s-length standard�when
determining transfer prices used by related parties. Under the arm�s-length standard,
1
the transfer price should be the price at which two unrelated parties engaging in a
comparable transaction under comparable circumstances would trade. There are
various methods that can be used when applying this principle in practice, including
the comparable uncontrolled transaction method, the resale price method, the
cost-plus method, the comparable pro�t method, and the pro�t-split method.
Furthermore, how these methods are applied in practice depends in part on which
transactions between unrelated parties are considered to be most comparable to the
related party transaction. If when faced with these various methods the two countries
would derive the same transfer price, we consider the transfer price rules to be
consistent. If the two countries would derive di¤erent prices, we consider the transfer
price rules to be inconsistent.
Inconsistent application of the arm�s-length standard can cause the aggregate
taxable income in the two countries to exceed the worldwide accounting income from
the transaction, but cannot cause aggregate taxable income to be less than worldwide
pretax accounting income. This asymmetry in outcomes re�ects the asymmetric
position of the taxpayer and the tax authorities. The taxpayer must choose a single
transfer price that it uses when �ling both its domestic and foreign tax returns. Both
the domestic and foreign tax authorities can challenge this transfer price; a successful
challenge does not require the other tax authority to accept the new transfer price. A
consequence of the asymmetry in the positions of the taxpayer and the tax authorities
is that aggregate taxable income (after audit) from the transaction can exceed
2
worldwide pretax economic income from the transaction, but cannot be less.
Mechanisms such as mutual agreement procedures (MAPs) exist to make it
possible for a taxpayer to get relief from double taxation. However, in practice the
MAP system often fails to provide such relief because governments are not required to
resolve the con�ict in a manner that eliminates double taxation (Mortier 2002). In the
1997, 1999, and 2001 Ernst & Young LLP Transfer Pricing Global surveys, over 80% of
multinationals identi�ed �double tax relief�as an important international tax issue
because transfer pricing adjustments typically result in double taxation (Ackerman
and Hobster 2001). Transfer pricing continues to be the top international tax issue to
multinationals in the most recent survey, conducted in 2007 (Ernst and Young 2008).
In this paper, we investigate the e¤ect of transfer price inconsistency on
production location and productive e¢ ciency. In our model, productive e¢ ciency is
achieved if domestic and foreign investment face the same tax rate, i.e., if capital
export neutrality holds. Although e¢ ciency and capital export neutrality are not
equivalent in general [Horst 1980; Desai and Hines 2004], our focus is not on capital
export neutrality per se but rather on how transfer price inconsistency and
repatriation taxes a¤ect the location and e¢ ciency of production.
We �rst examine a benchmark case that corresponds to the way that the U.S.
taxes foreign income, in which worldwide income is subject to tax by the U.S.
government, a credit for foreign taxes paid (subject to limitations) is available,
domestic taxation of foreign source income is deferred until repatriation, and double
3
taxation due to inconsistent transfer prices is possible. We �nd that the current tax
system can induce the e¢ cient level of investment, excessive domestic investment, or
excessive foreign investment, depending on the tax rates and the extent of transfer
price inconsistency. Eliminating double taxation by harmonizing transfer prices either
has no e¤ect on investment or increases foreign investment; in the latter case, the shift
towards foreign investment could either increase or decrease e¢ ciency. The e¤ects of
reductions in either the domestic or foreign corporate tax rate, or changes in the
shareholder after-tax discount rate, depend in part on the extent of transfer price
inconsistency.
In Section 2 we review the relevant theoretical and empirical literature. We
present our model in Section 3. In Section 4, we examine a benchmark case that
re�ects the current U.S. tax law. In Section 5 we investigate the e¤ects of harmonizing
transfer pricing rules on the location and e¢ ciency of investment. Section 6
investigates the e¤ect of lowering domestic and foreign corporate tax rates, and of a
change in the shareholder after-tax discount rate, on investment location and
e¢ ciency. Section 7 concludes.
2 Prior literature
The model we use to investigate the e¤ects of inconsistent transfer pricing rules was
developed by Anand and Sansing (2000) in their study of formulary apportionment,
4
the system used by U.S. states and Canadian provinces to allocate taxable income
among political jurisdictions. De Waegenaere and Sansing (2008b) also use this model
to compare separate accounting to formulary apportionment.
Most papers that examine tax transfer pricing rules assume that all political
jurisdictions use the same transfer price; however, there are several exceptions. De
Waegenaere, Sansing, and Wielhouwer (2006, 2007) examine inconsistent transfer
pricing rules in the context of strategic tax compliance models. In those papers,
however, the production decisions preceding the tax compliance decision are taken as
�xed. Halperin and Srinidhi (1987) and Elitzur and Mintz (1996) study the e¤ect of
inconsistent tax transfer pricing rules on production decisions under conditions of
imperfect competition. As in this paper, De Waegenaere and Sansing (2008b) focus on
the production decisions that inconsistent transfer pricing rules induce. That paper
focuses on the e¤ects of changing from a system of separate accounting to formulary
apportionment; this paper focuses on the e¤ects of harmonizing transfer prices, and on
how the e¤ects of changes in corporate tax rates and in the shareholder after-tax
discount rate depend on the level of transfer price inconsistency.
Our paper also relates to the literature on the e¤ect of taxes on repatriation
decisions. Hartman (1985) was the �rst to examine the e¤ect of the U.S. tax system on
foreign investment and repatriation decisions. He showed that the U.S. repatriation
tax does not distort the decision by the foreign subsidiary whether to reinvest its
earnings in a new project or pay a dividend, because all earnings are reduced by the
5
same repatriation tax rate sooner or later. Hines and Rice (1994), Weichenrieder
(1996), and Altshuler and Grubert (2002) extend Hartman�s model to consider the role
of investment in �nancial assets by a foreign subsidiary. These papers show that
investing in risk-free �nancial assets costlessly defers the repatriation tax when the
parent corporation discounts riskless after-tax cash �ows using the after-domestic
corporate tax risk-free rate. Therefore, in these models, the subsidiary should reinvest
earnings on operating assets in the risk-free asset instead of repatriating them to the
parent as a dividend. Using Brennan�s (1970) after-tax capital asset pricing model, De
Waegenaere and Sansing (2008a) argue that the appropriate discount rate should
re�ect the average shareholder tax rate and not the domestic corporate tax rate. In
their model, investing in the risk-free asset to avoid the repatriation tax involves an
opportunity cost to the extent the after-domestic corporate tax interest rate is less
than the after-tax discount rate.
3 Model
We study an economy with one good and two countries. Demand for this good is
perfectly inelastic. Domestic demand for the good is �; foreign demand for the good is
zero. Because the focus of our study is on how inconsistent transfer prices and tax
rates a¤ect the location of production, it is convenient to assume that total demand
for (and thus the aggregate equilibrium supply of) the good is �xed, and focus on
6
where production takes place.
Production can occur in either country, employing a constant returns to scale
production technology with a single non-depreciable input that we call capital. One
unit of physical capital is needed to produce one unit of output. We assume that some
inputs are immobile (it may be useful to think of land or natural resources as being an
important component of capital), which implies that each country has its own supply
curve. The price per unit of physical capital associated with production in each
country is determined by aggregate production in that country (denoted qi), according
to
Ci(qi) = �iqi: (1)
The supply curve in each country is upward-sloping, and Ci(0) = 0: The
upward-sloping supply curves imply that the competing use for physical capital varies
within each country; the assumption that Ci(0) = 0 ensures that some production
occurs in each country.
All sales occur in the domestic country. One unit of non-depreciable
physical capital is needed to sell one unit. The price per unit of capital associated with
sales is K: A �rm that produces domestically and sells one unit incurs a cost of capital
of r[CD(qD) +K], where r denotes the cost of equity capital. A foreign subsidiary that
produces one unit incurs a cost of capital of rCF (qF ), whereas its domestic parent
incurs a cost of capital of rK:
There are many small private �rms incorporated in the domestic country
7
engaged in production under conditions of perfect competition, in the sense that each
�rm takes input costs and output prices as given. Each �rm can produce domestically
or in the foreign country. We let p denote the equilibrium price of the good.
Taxes
Domestic production is taxed by the domestic government at a rate of �D < 50% on its
pretax accounting income each period, i.e., exclusive of its cost of capital, so a �rm
that produces domestically and sells one unit has after-tax accounting income of
p(1� �D). The income of p from foreign production is subject to tax by both the
foreign and domestic governments. The allocation of income for tax purposes between
the two countries depends on the transfer price. We let �F (�D) denote the fraction of
income associated with foreign production that according to the transfer price used for
tax purposes by the foreign government (domestic government) should be taxed by the
foreign government, where
0 � �D � �F � 1:
The upper and lower bounds ensure that taxable income in each country is weakly
positive. We let
I =�F�D
� 1
represent the degree of transfer price inconsistency between the two countries. When
I = 1; the transfer prices are consistent and so the taxable income in the foreign
8
country is equal to the tax deduction permitted by the domestic government; when
I > 1; the taxable income in the foreign country exceeds the tax deduction allowed in
the domestic country, causing more than 100 percent of the income from the
transaction to be taxed.
We let �F denote the foreign country�s corporate tax rate, and assume that
�F < 50%. Producing one unit in the foreign country generates after-tax domestic
accounting income of p(1� �D)(1� �D), and after-foreign tax foreign earnings and
pro�ts of p (�D � �F�F ) : The domestic government uses a worldwide tax system, in
which income earned in the foreign country is subject to tax by the domestic
government when after-foreign tax earnings are repatriated to the domestic parent via
a dividend. The domestic government allows a credit for foreign taxes, but not to
exceed the domestic tax on foreign source income. Three cases are possible. In the
�rst case, I�F � �D; so no repatriation tax is due because the foreign tax exceeds the
domestic tax on foreign source income (�F�F � �D�D). Note that this occurs either
when �D � �F ; so that the foreign tax exceeds the domestic tax on foreign source
income, regardless of the transfer pricing rules, or when �F < �D and I � �D�F; in which
case the inconsistent transfer prices cause the foreign tax to exceed the domestic tax,
even though the domestic tax rate exceeds the foreign tax rate. The after-tax
accounting pro�t from producing one unit in the foreign country is then given by
p(1� �D)(1� �D) + p [�D � �F�F ] = p [1� �D(1� �D)� �F�F ] : (2)
In the second and third cases, I�F < �D, and the foreign tax is lower than the
9
domestic tax on foreign source income (�F�F < �D�D), so that the repatriation of
after-foreign tax earnings would trigger a repatriation tax of �D�D � �F�F : The
repatriation tax transforms the after-foreign tax earnings of �D � �F�F into an
after-repatriation tax amount of �D(1� �D): If the foreign subsidiary repatriates its
after-foreign tax earnings, the �rm�s after-tax income equals
p(1� �D): (3)
However, the foreign subsidiary can instead invest the after-foreign tax earnings in
bonds that earn the risk-free rate of R, thereby inde�nitely deferring the repatriation
tax. The interest is taxed at the domestic rate �D; of which �F is collected by the
foreign government and �D � �F is collected by the domestic government. Because the
interest is taxed immediately under Subpart F, there is no reason to defer repatriating
the interest income. De Waegenaere and Sansing (2008a) show that Brennan�s (1970)
after-tax capital asset pricing model implies that the appropriate discount rate to use
when determining the present value of the future interest income earned on the
after-foreign tax earnings is R(1� �S), where �S re�ects the average shareholder tax
rate on interest; we present their argument for the use of R(1� �S) to discount riskless
after-tax cash �ows in Appendix A. We assume that �S � �D. Thus the present value
to the domestic parent of the after-foreign tax earnings is (�D��F�F )R(1��D)R(1��S) , and the
present value of the future after-tax cash �ows assocated with current sale of one unit
10
equals
p
�(1� �D)(1� �D) +
(�D � �F�F )(1� �D)1� �S
�= p
�1� �D +
(�S�D � �F�F )(1� �D)1� �S
�: (4)
Comparing the after-tax payo¤s to the domestic parent from (3) and (4) shows that
the foreign subsidiary will immediately repatriate earnings if �F�F � �S�D, i.e. if
I�F � �S, and invest in �nancial assets otherwise. The intuition is as follows. When
�F�F � �D�D, a domestic tax is due upon repatriation. The repatriation tax can be
avoided by having the foreign subsidiary invest its after-foreign tax earnings in
�nancial assets. However, there is an opportunity cost associated with investing in
�nancial assets, because �S � �D implies that the after-tax return on the bond is
weakly lower than the after-tax discount rate for riskless cash �ows. As long as
�F�F � �S�D, the bene�t of avoiding the repatriation tax exceeds the opportunity
cost of investing in �nancial assets.
The above analysis shows that, in each case, the after-tax pro�t from foreign
production is of the form p(1� T ), where T re�ects the combined e¤ect of foreign and
domestic taxes imposed on income generated by foreign production, plus the
opportunity cost associated with investment in �nancial assets. The following
proposition summarizes the e¤ect of the tax system, tax rates, and transfer prices on
the combined tax rate T imposed on foreign production.
11
Proposition 1 (a) If I�F > �D, then there is no repatriation tax, and
T = �D + �F �F � �D�D > �D;
(b) if �S � I�F � �D; then the repatriation tax causes all income to be taxed at the
domestic country�s tax rate, so T = �D;
(c) if I�F < �S, then the repatriation tax is avoided via investment in �nancial
assets, and T = �D � (�S�D��F�F )(1��D)1��S < �D:
Three e¤ects determine T : the relation between the corporate tax rates �D and
�F ; the relation between the foreign tax rate and the shareholder tax parameter �S;
and the degree of transfer price inconsistency, I: First, suppose there is no transfer
price inconsistency, i.e. �F = �D, and so I = 1: Then T > �D if �F > �D because the
foreign tax on foreign income exceeds the domestic tax on foreign income; there is no
repatriation tax in that case: T < �D if �F < �S because then the cost of avoiding the
repatriation tax by investing foreign earnings in �nancial assets is su¢ ciently low. The
critical comparison is between �F and �S, and does not involve �D, because
repatriating one dollar of after-foreign tax earnings yields an after-repatriation tax cash
�ow of 1��D1��F ; whereas reinvesting that same dollar in �nancial assets and repatriating
the interest yields an after-repatriation tax perpetual cash �ow of R(1� �D) dollars
that has a present value of 1��D1��S dollars. Finally, T = �D if �S � �F � �D because
then the cost of avoiding the repatriation tax is too high: To illustrate the e¤ect of
transfer price inconsistency, consider the case in which �F < �S < �D: Then transfer
12
price inconsistency can imply that T is greater than, equal to, or less than �D.
Speci�cally, T > �D if transfer price inconsistency is su¢ ciently high (I > �D�F); T < �D
if transfer price inconsistency is su¢ ciently low (I < �S�F); and T = �D if �S�F � I �
�D�F:
Equilibrium
We de�ne a competitive equilibrium to be an output price p and aggregate output
quantities qD and qF at which both domestic and foreign production earns zero
after-tax economic pro�ts. In every case, this requires for domestic production
p(1� �D) = rCD(qD) = r(�DqD +K): (5)
The equilibrium condition for foreign production is given by
p(1� T ) = rCF (qF ) = r(�F qF +K); (6)
where T depends on the subsidiary�s repatriation tax and repatriation strategy, as
given in Proposition 1. In addition, total output must equal demand, so
qD + qF = �: (7)
The equilibrium price and quantities follow from solving the set of equations (5)-(7).
Without taxes, equilibrium production is e¢ cient and given by qD = ��F�D+�F
and
qF =��D
�D+�F: With taxes, the equilibrium price is given by
p =r[��D�F +K(�D + �F )]
�D(1� T ) + �F (1� �D);
13
and the equilibrium production quantities qD and qF are given by
qD =��F +K
�1� 1�T
1��D
��D
�1�T1��D
�+ �F
; qF =��D +K
�1� 1��D
1�T�
�D + �F�1��D1�T
� : (8)
We assume that � is su¢ ciently large to ensure that all prices and quantities are
positive.1 It follows from (8) that the e¤ects of the tax system, tax rates, and transfer
prices on equilibrium production decisions are determined by
� =1� T1� �D
;
the ratio of the after-tax accounting pro�t from foreign production to the after-tax
accounting pro�t from domestic production. The fraction of demand satis�ed by
domestic production is strictly decreasing in �.
Productive e¢ ciency
The e¢ cient outcome is the pair (qD; qF ) that minimizes the cost of producing � units
of output; the cost of selling � units is r�K; and does not depend on where production
occurs. To �nd this pair, we �rst use (1) to show that the social cost of producing qi
units in country i isqiZ0
r�iqidqi =r�iq
2i
2; for i 2 fD;Fg: (9)
Next, we solve the following cost minimization problem:
minqD;qF
nr�Dq
2D+r�F q
2F
2
os:t: qD + qF = �;
(10)
1Our assumptions that �F < 50% and �D < 50% jointly ensure that T < 1, so that such a � exists.
14
the solution to which is
qD =��F
�D + �F; qF =
��D�D + �F
: (11)
It follows from (8) and (11) that production is e¢ cient if T = �D, ine¢ cient with
excess production in the domestic country if T > �D, and ine¢ cient with excess
production in the foreign country if T < �D. Therefore, this model re�ects capital
export neutrality in that productive e¢ ciency is achieved if and only if domestic and
foreign investment face the same tax rate.
4 Production decisions under the status quo
In this section, we use Proposition 1 to investigate the equilibrium production
decisions under the current U.S. tax rules that feature worldwide taxation, deferral of
domestic taxation of foreign earnings until repatriation, and inconsistent transfer
prices. This provides a benchmark case with which we then will consider possible tax
changes in transfer prices or tax rates.
Proposition 1 implies that the combined tax rate imposed on foreign production
depends on the degree of transfer price inconsistency I. Depending on tax rates and
transfer prices, the after-tax accounting pro�t from foreign production can be either
higher or lower than the after-tax accounting pro�t from domestic production under
both systems.
When transfer price inconsistency is high, I�F > �D, Proposition 1(a) shows that
15
T > �D: Therefore, production is ine¢ cient with excessive domestic investment. When
transfer price inconsistency is moderate, �S � I�F � �D; Proposition 1(b) shows that
T = �D; and so production is e¢ cient. When transfer price inconsistency is low,
I�F < �S; Proposition 1(c) shows that T < �D: Therefore, production is ine¢ cient
with excessive foreign investment. We summarize these results in Table 1.
[INSERT TABLE 1 ABOUT HERE]
5 Harmonizing transfer prices
In this section, we investigate the e¤ects of harmonizing transfer pricing rules on
production decisions and productive e¢ ciency. We consider a setting in which transfer
price inconsistencies are eliminated. Speci�cally, both countries agree that the fraction
of a foreign producer�s income taxed by the foreign country is given by �H ; where
�D � �H � �F :
We let TH (T ) denote the combined tax rate imposed on foreign production in case of
consistent (inconsistent) transfer prices. We use Proposition 1 to determine T and TH ,
and then compare (8) to (11) to investigate the e¤ect of harmonizing transfer prices on
production location and productive e¢ ciency.
Harmonization of transfer prices implies that the domestic tax on domestic
income from foreign production decreases from (1� �D)�D to (1� �H)�D, and the
foreign tax on foreign income decreases from �F �F to �H�F . The e¤ect on TH ,
16
however, depends also on how harmonization a¤ects the repatriation tax and
repatriation decisions.
It follows from Proposition 1 that with harmonized transfer prices (i.e. I = 1),
the combined tax rate imposed on foreign production depends on the tax rates �F ; �D;
and �S. With inconsistent transfer prices, production decisions also depend on the
degree of transfer price inconsistency. Because I � 1, six cases are possible.
Case 1: �F � �D. Then, both with consistent and inconsistent transfer prices, the
foreign tax on foreign income exceeds the domestic tax on foreign income. Proposition
1 shows that
TH = �D + �H (�F � �D) � �D: (12)
T = �D + �F �F � �D�D � �D: (13)
Output is ine¢ cient with excessive domestic production in both cases. Subtracting T
from TH yields
TH � T = �F (�H � �F ) + �D(�D � �H) � 0:
Because harmonization decreases both the foreign and the domestic tax on foreign
income, and there is no repatriation tax, it decreases the combined tax imposed on
foreign production. Therefore, harmonizing transfer prices increases foreign production
and increases e¢ ciency in this case.
Case 2: �S � �F � �D and �D � I�F . Then, whereas under inconsistent transfer
prices the foreign tax on foreign income exceeds the domestic tax, the opposite holds
17
under consistent transfer prices. When transfer prices are inconsistent, Proposition
1(a) shows that T � �D, and so production is ine¢ cient with excessive domestic
production. If transfer prices are harmonized, I = 1, and Proposition 1(b) shows that
TH = �D; which implies that the e¢ cient level of production is achieved. So in this
case, harmonizing transfer prices increase foreign production and yields the e¢ cient
outcome.
Case 3: �F � �S and �D � I�F . Then, as in Case 2, harmonization brings the foreign
tax on foreign income down from above the domestic tax to below the domestic tax.
However, Proposition 1(c) with I = 1 now implies that deferring the repatriation tax
through investment in �nancial assets is optimal with consistent transfer prices.
Therefore, although T � �D, harmonization implies that TH � �D: So in this case,
harmonizing transfer prices increases foreign production, inducing a shift from
ine¢ cient excessive domestic production to ine¢ cient excessive foreign production.
The overall e¤ect on e¢ ciency is ambiguous.
Case 4: �S � �F � �D and �S � I�F � �D. In this case, the foreign tax on foreign
income is lower than the domestic tax, even with inconsistent transfer prices.
Proposition 1(b) shows that with either consistent or inconsistent transfer prices,
immediately repatriating income is preferred to investing in �nancial assets, and so
T = �D and TH = �D. Productive e¢ ciency is achieved in both cases, and harmonizing
transfer prices has no e¤ect on production decisions.
Case 5: �F � �S and �S � I�F � �D. As in Case 4, the foreign tax on foreign income
18
is lower than the domestic tax, even under inconsistent transfer prices. However, in
this case it follows from Proposition 1(b) and Proposition 1(c) with I = 1 that,
whereas under inconsistent transfer prices it is optimal to immediately repatriate
income, yielding T = �D, under harmonized transfer prices, it is optimal to avoid the
repatriation tax through investment in �nancial assets, yielding TH � �D, and there is
excessive foreign production. Therefore, harmonizing transfer prices increases foreign
production and reduces e¢ ciency in this case.
Case 6: �F � �S and I�F � �S. In this case, Proposition 1(c) shows that avoiding the
repatriation tax by investing in �nancial assets is optimal both under inconsistent
transfer prices and under harmonized transfer prices. Speci�cally,
TH = �D � �H(�S � �F )(1� �D)
1� �S� �D (14)
T = �D �(�D�S � �F �F )(1� �D)
1� �S� �D: (15)
In both cases, production is ine¢ cient with excessive foreign production. Subtracting
T from TH shows that
TH � T = �F (�H � �F ) + �S(�D � �H) � 0:
Therefore, transfer price harmonization induces greater foreign investment and
decreases e¢ ciency in this case.
The above six cases show that transfer price harmonization can a¤ect both the
location and e¢ ciency of production, and that both these e¤ects depend on whether
foreign income is heavily taxed relative to domestic income (i.e. I�F vs. �D), and on
19
whether �S is lower or higher than �F . Harmonization increases foreign production
when foreign income is heavily taxed relative to domestic income (I�F � �D):
Harmonization also increases foreign production when foreign income is lightly taxed
relative to domestic income (I�F � �D) and the opportunity cost of deferring the
repatriation tax by having the foreign subsidiary invest in �nancial assets is su¢ ciently
low (�S � �F ): When foreign income is lightly taxed relative to domestic income, but
the opportunity cost of deferring the repatriation tax by having the foreign subsidiary
invest in �nancial assets is su¢ ciently high (�S < �F ), harmonization has no e¤ect on
the location of production.
Transfer price harmonization has ambiguous e¤ects on productive e¢ ciency.
When foreign income is heavily taxed, harmonization increases e¢ ciency if �F � �S by
bringing the combined tax rate on foreign income closer to the domestic tax rate.
When �F < �S, however, harmonization brings the combined tax rate on foreign
income down below the domestic tax rate, yielding excessive foreign investment; this
induces a shift from ine¢ ciently high to ine¢ ciently low domestic production. The
overall e¤ect of harmonization could increase or decrease e¢ ciency in that case. When
foreign income is lightly taxed, harmonization decreases e¢ ciency if �S > �F and has
no e¤ect if �F � �S. We summarize the e¤ect of transfer price harmonization on the
location and e¢ ciency of production in Table 2.
[INSERT TABLE 2 ABOUT HERE]
20
6 Lowering tax rates
In this section, we use Proposition 1 to investigate how the e¤ects of domestic and
foreign corporate tax rate reductions and changes in the shareholder after-tax discount
rate depend on the degree of transfer price inconsistency.
Domestic corporate tax rate
The �rst tax rate change we examine involves lowering the corporate statutory tax
rate �D. The U.S. currently has one of the highest tax rates among OECD nations,
which as we have seen can induce excessive foreign investment. We analyze the e¤ect
of a small decrease in �D on the location and e¢ ciency of investment. A change in the
domestic tax rate a¤ects both the tax rate on domestic production and the combined
tax rate T on foreign production. It follows from (8) that the e¤ect of a change in the
domestic tax rate on equilibrium production decisions depends on how it a¤ects the
ratio � = 1�T1��D of the after-tax pro�t from foreign production to the after-tax pro�t
from domestic production. A change in �D increases the fraction of demand satis�ed
by domestic production if it decreases �. It follows from (8) and (11) that a change in
�D increases e¢ ciency if it drives the ratio � closer to one. We emphasize that e¤ects
on production location and productive e¢ ciency depend on �, and not on the
di¤erence between the tax rates T and �D.
We consider two cases. If I�F > �D; then T > �D and thus foreign production
has a lower after-tax pro�t than domestic production, and so � < 1: Proposition 1(a)
21
implies that di¤erentiating � with respect to �D yields
@�
@�D=�D � �F �F(1� �D)2
;
which could be positive or negative, depending on whether I�F > 1 or I�F < 1: If
I�F > 1, then @�@�D
< 0, and so decreasing the domestic corporate tax rate increases �,
thereby increasing foreign investment and increasing e¢ ciency. If I�F < 1, then
@�@�D
> 0, and so decreasing the domestic corporate tax rate decreases �, thereby
increasing domestic investment and decreasing e¢ ciency. The e¤ect of a tax rate
decrease depends on the level of transfer price inconsistency.
In contrast, a small decrease in �D has no e¤ect on production decisions if
I�F � �D. Indeed, if �S � I�F � �D or I�F < �S, then parts (b) and (c) of
Proposition 1 imply that the ratio � of after-tax pro�ts of foreign and domestic
production is independent of �D: Therefore, a change in �D has no e¤ect on
production decisions in these cases. If �S � I�F � �D; a change in �D has no e¤ect
because the repatriation tax implies that foreign income and domestic income are
taxed at the same tax rate. If I�F < �S, a change in �D has no e¤ect because both the
income that is allocated to the domestic parent when it is earned (p(1� �D)) and the
interest from after-foreign tax foreign earnings (R(�D � �F�F )) face a tax rate of �D.
We summarize the results of decreasing the domestic corporate tax rate �D in
Table 3. If foreign income is heavily taxed relative to domestic income, I�F > �D;
decreasing �D decreases e¢ ciency if transfer prices are very inconsistent (I�F > 1) and
increases e¢ ciency if transfer price inconsistency is not too extreme (I�F < 1): If
22
foreign income is lightly taxed relative to domestic income, decreasing �D has no e¤ect
on the location of investment and no e¤ect on productive e¢ ciency.
[INSERT TABLE 3 ABOUT HERE]
Foreign corporate tax rate
Next, we examine the e¤ect of lowering the foreign corporate statutory tax rate �F .
Corporate statutory tax rates have been steadily declining since 2000 (Ernst and
Young 2007). We analyze the e¤ect of a small decrease in �F on the location and
e¢ ciency of investment. In contrast to the case of a domestic tax rate cut, a change in
the foreign tax rate a¤ects only the combined tax rate T on foreign production. It
increases the fraction of demand satis�ed by domestic production if it increases T . A
change in �F increases e¢ ciency if it drives T closer to �D.
We again consider the three cases in Proposition 1. If I�F � �D, then T > �D.
Because a small decrease in �F decreases the foreign tax on foreign income, and there
is no repatriation tax, it decreases T , thereby increasing foreign investment and
increasing e¢ ciency. If �S � I�F � �D; then investment decisions are e¢ cient
(T = �D), and a small change in �F has no e¤ect on investment decisions or e¢ ciency.
Finally, if I�F < �S, then T < �D. Because a small decrease in �F decreases the
foreign tax on foreign income, and the burden of the repatriation tax can be partly
mitigated through investment in �nancial assets, a cut in the foreign tax rate decreases
T . Consequently, it increases foreign investment, and decreases e¢ ciency.
23
We summarize the results of decreasing the foreign corporate tax rate �F in
Table 4. Because a small decrease in �F decreases the foreign tax on foreign income, it
increases foreign production unless �S � I�F � �D, in which case the repatriation tax
implies that all income is taxed at the rate �D. However, the e¤ect on e¢ ciency is
ambiguous. If foreign income is heavily taxed relative to domestic income, I�F > �D;
decreasing �F increases e¢ ciency by bringing the combined tax rate on foreign
investment closer to the domestic tax rate. If foreign income is lightly taxed relative to
domestic income, I�F � �D, the e¤ect of decreasing �F further depends on a
comparison between I�F and �S, increasing an already ine¢ ciently high level of
foreign investment if I�F < �S and having no e¤ect otherwise.
[INSERT TABLE 4 ABOUT HERE]
Shareholder after-tax discount rate
Finally, we examine the e¤ect of changes in the after-tax discount rate used by
shareholders to value the �rm�s riskless after-tax cash �ows on the location and
e¢ ciency of production undertaken by multinational �rms. As shown in the appendix,
the discount rate is R(1� �S), where R is the risk-free interest rate and the parameter
�S re�ects the average shareholder tax rate on interest; �S is increasing in the tax rate
that taxable investors face on bond interest and in the fraction of taxable investors in
the economy. We focus on the parameter �S, a decrease in which increases the
after-tax discount rate investors use when valuing riskless cash �ows. This in turn
24
determines whether deferring the repatriation tax on foreign earnings by having the
foreign subsidiary invest in �nancial assets increases �rm value.
As in the case of a foreign tax rate cut, a change in the shareholder tax rate
a¤ects only the combined tax rate T on foreign production, and the e¤ect on
production location and productive e¢ ciency depends on whether the change reduces
the di¤erence between the tax rates imposed on foreign and domestic production.
There are two cases to consider. If �S � I�F ; then �S is so low that deferring the
repatriation tax via investment in �nancial assets is unattractive, so a further
reduction would have no e¤ect on production decisions or e¢ ciency. If �S > I�F , then
Proposition 1(c) shows that T < �D. A decrease in �S increases T; because it increases
the cost of avoiding the repatriation tax. Therefore, it increases domestic investment
and increases e¢ ciency. We summarize the results of increasing the shareholder
after-tax discount rate by decreasing the shareholder tax parameter �S in Table 5.
[INSERT TABLE 5 ABOUT HERE]
7 Conclusions
We have examined the e¤ects of inconsistent transfer prices on the location of capital
investments in a competitive equilibrium. The decision to invest domestically or in a
foreign country depends jointly on statutory tax rates, the extent to which transfer
prices are inconsistent, and the after-tax discount rate shareholders use to value
25
riskless cash �ows. We also examine the e¤ects of transfer price inconsistency on
productive e¢ ciency in a model in which productive e¢ ciency is equivalent to capital
export neutrality. Eliminating the di¤erence between domestic and foreign transfer
pricing rules makes foreign investment weakly more attractive, but could either
increase or decrease productive e¢ ciency. The e¢ ciency e¤ects of reductions in either
the domestic or the foreign corporate tax rates, or changes in the shareholder after-tax
discount rate, also depend on the degree of transfer price inconsistency.
26
References
Ackerman, R. and J. Hobster. 2001. Transfer pricing practices, perspectives, and
trends in 22 countries. Tax Notes International 24 (December 17): 1151-1158.
Altshuler, R., and H. Grubert. 2002. Repatriation taxes, repatriation strategies
and multinational �nancial policy. Journal of Public Economics 87 (January): 73-107.
Anand, B. and R. Sansing. 2000. The weighting game: Formula apportionment
as an instrument of public policy. National Tax Journal 53 (June): 183-99.
Brennan, M. 1970, Taxes, market valuation and corporate �nancial policy,
National Tax Journal 23: 417-427.
De Waegenaere, A., R. Sansing, and J. Wielhouwer. 2006. Who bene�ts from
inconsistent multinational tax transfer pricing rules? Contemporary Accounting
Research 23 (Spring): 103-131.
De Waegenaere, A., R. Sansing, and J. Wielhouwer. 2007. Using bilateral
advance pricing agreements to resolve tax transfer pricing disputes. National Tax
Journal 60 (June): 173-191.
De Waegenaere, A. and R. Sansing. 2008a. Taxation of international investment
and accounting valuation. Contemporary Accounting Research 25 (Winter):
forthcoming.
De Waegenaere, A. and R. Sansing. 2008b. Taxing multinational �rms: transfer
pricing or formulary apportionment? Working paper.
27
Desai, M. and J. Hines. 2004. Old rules and new realities: corporate tax policy
in a global setting. National Tax Journal 57 (December): 937-60.
Ernst & Young. 2007. Tax administration goes global: complexity, risks and
opportunities. Tax Notes International 45 (February 26): 813-27.
Ernst & Young. 2008. Global transfer pricing trends, practices, and analyses.
Tax Notes 118 (February 18): 845-60.
Elitzur, R. and J. Mintz. 1996. Transfer pricing rules and corporate tax
competition. Journal of Public Economics 60 (June): 401-422.
Halperin, R., and B. Srinidhi. 1987. The e¤ects of U.S. income tax regulations�
transfer pricing rules on allocative e¢ ciency. The Accounting Review 62 (October):
686-706.
Hartman, D. 1985. Tax policy and foreign direct investment. Journal of Public
Economics 26 (February): 107-121.
Hines, J. 1994. Credit and deferral as international investment incentives.
Journal of Public Economics 55 (October): 323-47.
Hines, J. and E. Rice. 1994. Fiscal paradise: foreign tax havens and American
business. Quarterly Journal of Economics 109 (February): 149-182.
Horst, T. 1980. A note on the optimal taxation of international investment
income. Quarterly Journal of Economics 94 (June): 793-98.
Mortier, F. 2002. International tax arbitration: Toward better taxpayer
protection. Tax Notes International 27 (September 30): 53-58.
28
Weichenreider, A. 1996. Anti tax-avoidance provisions and the size of foreign
direct investment. International Tax and Public Finance 3 (January): 67-81.
29
Investment e¢ ciency under the status quo
E¢ ciency
�D < I�F Excess domestic investment
�S � I�F � �D E¢ cient investment
I�F < �S Excess foreign investment
Table 1
E¤ect of harmonzing transfer prices on investment location and e¢ ciency
Production location E¢ ciency
I�F � �D; �S � �F Increase foreign investment Increase e¢ ciency
I�F � �D; �S � �F No e¤ect No e¤ect
I�F � �D; �F � �S Increase foreign investment Ambiguous e¤ects on e¢ ciency
I�F � �D; �F � �S Increase foreign investment Decrease e¢ ciency
Table 2
30
E¢ ciency and location e¤ects of decreasing the domestic corporate tax rate
Production location E¢ ciency
1 � I�F Increase foreign investment Increase e¢ ciency
�D � I�F � 1 Increase domestic investment Decrease e¢ ciency
I�F � �D No e¤ect No e¤ect
Table 3
E¢ ciency and location e¤ects of decreasing the foreign corporate tax rate
Production location E¢ ciency
�D < I�F Increase foreign investment Increase e¢ ciency
�S � I�F � �D No e¤ect No e¤ect
I�F < �S Increase foreign investment Decrease e¢ ciency
Table 4
E¢ ciency e¤ects of increasing the shareholder after-tax discount rate
Production location E¢ ciency
�S � I�F No e¤ect No e¤ect
I�F � �S Increases domestic investment Increases e¢ ciency
Table 5
31
Appendix
We consider a model with M +N risk-averse investors who can invest in either a
riskless bond or a risky stock. The riskless bond pays interest at the rate R on each
date in perpetuity. The stock generates a risky return ex on each date in perpetuity.The risky return is normally distributed with a mean of � and a variance of �2. The
investors are of two types: taxable and tax-exempt. There are M taxable investors
that face a constant statutory tax rate tB on interest from the and a constant e¤ective
tax rate tG on the stock return. The e¤ective tax rate tG can be lower than the
statutory rate on realized capital gains because the tax on unrealized gains is deferred
until the stock is sold, and tax may be avoided altogether through a charitable gift of
the stock, or through a basis step-up upon the investor�s death. The remaining N
investors are tax-exempt.
Each investor has a utility function over end-of period wealth of the form
U(w) = �e��w. This utility function has the property that if ew is normallydistributed with mean � and variance �2, the investor�s certainty equivalent is equal to
�� ��2
2: We normalize the number of shares outstanding for the stock to be one.
An equilibrium is de�ned as a portfolio for each investor that maximizes that
investor�s expected utility given the stock price, and a stock price at which supply
equals demand. Each of the M taxable investors buys sT shares at a price of P per
share to solve the following maximization problem.
32
maxsT
�sT�(1� tG)� s2T��2(1� tG)2=2
R(1� tB)� sTP
�
Di¤erentiation yields the following �rst-order condition for each of the M taxable
investors.
s�T =�(1� tG)� PR(1� tB)
��2(1� tG)2
Each of the N tax-exempt investors buys sE shares of stock at a price of P per
share to solve the following maximization problem.
maxsE
�sE�� s2E��2=2
R� sEP
�
Di¤erentiation yields the following �rst-order condition for each of the N
tax-exempt investors.
s�E =�� PR��2
The market clearing condition completes the characterization of the equilibrium.
Ms�T +Ns�E = 1
Substituting each investor�s demand into the market clearing condition and
solving for P yields the equilibrium stock price.
P =�h
M1�tG +N
i� ��2
RhM(1�tB)(1�tG)2
+Ni
33
We determine the appropriate discount rate for riskless cash �ows by �nding the
discount rate rf for which��rf= �P , because an increase in � with no corresponding
increase in �2 represents an increase in riskless cash �ows. Therefore,
rf =1
@P=@�:
Di¤erentiating P with respect to � and solving for rf yields
rf =RhM(1�tB)(1�tG)2
+Ni
M1�tG +N
:
Finally, we express the discount rate as rf = R(1� tS) to highlight the relation
between investor tax rates and the appropriate discount rate. Solving for tS yields
tS =
M(tB�tG)(1�tG)2
M1�tG +N
:
When tG = 0; tS = MtBM+N
; the average shareholder tax rate on interest; when tG > 0;
tS <MtBM+N
: For the discount rate R(1� tD) to be normatively appropriate in our model,
we require N = 0; tG = 0; and tB = tD: In other words, no tax-exempt investors own
stock taxable investors face a tax rate on bond interest equal to the domestic corporate
tax rate, and taxable investors face a zero e¤ective tax rate on accrued capital gains.
34